【Proposal】Leveraging LLM-based Multi-Agent Collaboration to Enhance Embodied Agents’ Reasoning Capabilities for Solving Text-based Tasks in Human-populated Environments

Nan Sun, Chengming Shi, Yuwen Dong

Keywords: Multi-agent System, LLM-based Agent, Autonomous Robot, Human-robot Interaction, Embodied AI

TL;DR: A framework for leveraging LLM-based multi-agent collaboration to enhance embodied robots’ ability to execute text-based tasks in human-populated environments.

Abstract: This proposal explores the design of a reasoning framework leveraging LLM-based multi-agent collaboration to enhance the reasoning capabilities of embodied agents. By improving their understanding and execution of text-based instructions in complex, human-populated environments, the system aims to improve robots' dynamic reasoning, interaction with humans, and task completion. The proposed framework will enable robots to handle tasks autonomously while efficiently seeking human assistance when needed, ensuring task completion with minimal intervention.

Submission Number: 5